Anyone who’s ever watched that Friends episode knows the scene. Ross is wedged on a stairwell with a couch he has no business trying to move, screaming “PIVOT! PIVOT!” at Rachel and Chandler like volume is going to solve a geometry problem.
It’s funny because it’s useless advice delivered with total conviction. I hear this a lot lately, because “just pivot” is exactly the advice everyone gives you after fifty when you lose your job, and it’s about as helpful as Ross’s couch directions.
I was the oldest person on the team and the one who’d been there the longest. I was also one of the first to actually integrate AI into my daily workflow, not as a novelty, but as a real tool. I used it to cut down repetitive tasks that used to eat hours of my week. I used it to solve problems faster than I could have five years ago. None of that mattered in the end. There just wasn’t enough work left to need someone in my seat. It was a business decision. That’s it. That’s the whole reason. Not talent, not output, not attitude. Just less work to go around.
So here I am, wedged on my own stairwell, couch and all, trying to figure out which way to turn without putting anyone’s eye out.
Two very different doors
One door is teaching. I built out a full syllabus a while back, a real proposal for a college-level course on web design and development. I’ve also sketched a scaled-down version as a four session Saturday workshop, the kind of thing you could run for a mixed group with no real background, and have them walk away with an actual working WordPress site they built with their own hands.
The question I keep circling is whether any of that still matters when a fifteen year old with no training can ask AI to “build me a website” and get something that loads. I don’t think the answer is no. I think the answer is that the content of what gets taught has to shift, but the need for someone who actually understands what’s happening under the hood hasn’t gone anywhere. If anything it’s gotten sharper. AI tools are only as good as the person steering them. Someone who doesn’t understand the difference between a template hierarchy and a hook, or what a breakpoint actually does, will get garbage out of AI and not know why. The people who get real value out of these tools are the people who can read the output critically. That’s a teachable skill. It might be the only teachable skill left in this field that’s actually future-proof.
The other door is painting. Less about income, more about getting my hands on something physical and slow after three decades of pixels and deadlines. I’m not under any illusion it pays the bills. But I notice I’m drawn to it more as of late, and I think the reason is that I need something that isn’t graded by an algorithm or a client’s inbox.
The mentor problem
Here’s the part that actually gets me. The conventional wisdom for people my age is: don’t compete with the twenty five year olds on raw output, compete on judgment. Be the mentor. Be the senior voice who catches the mistakes before they go to production. That’s supposed to be our lane.
But mentor who? I’m watching entry level and junior developer roles get quietly erased. Companies are loading more onto a single senior dev and propping up the rest with AI agents that can grind through tickets all day without health insurance or a learning curve. If the junior tier disappears, the mentorship pitch falls apart. You can’t be valued as the person who trains up the next generation if there’s no next generation on the organization chart anymore. The ladder isn’t just harder to climb, the bottom rungs are getting sawed off.
And then there’s the hiring math
On top of that, I’m seeing fewer companies willing to take a chance on someone my age, even with a track record of staying put and doing the work, and doing it well. The math they’re running, intentionally or not, favors the twenty four year old who might leave in eight months over the fifty plus year old who’s shown up reliably for years. Loyalty and longevity used to be selling points on a resume. Now they read almost like a liability, because the cost structure rewards churn, not continuity. It’s a strange inversion. The thing that used to make you a safe bet now makes you look expensive and set in your ways, even when you’re the one in the room who adapted fastest.
That’s the part that stings the most, honestly. I wasn’t the guy dragging his feet on new tools. I was first to be all in. And it still wasn’t enough, because the decision wasn’t about me at all. It was about headcount and revenue, and the effect AI has had on our industry.
The numbers game
So I do what everyone tells you to do. I apply. I applied for unemployment for the first time ever. I reply to about three job postings a day, every day, which at this point feels less like a job search and more like a part time job with no pay and a brutal rejection rate. I’ve landed a handful of interviews out of it, which I’m choosing to count as a win, because a handful is more than zero and at this stage I’ll take encouraging over discouraging.
What’s actually landed me paid work, though, has nothing to do with any of those applications. It’s a contract gig through a friend, a few hours a week, that’s picked up a little recently. Which confirms the thing everyone already half knows and nobody likes saying out loud: it is still, overwhelmingly, who you know. The algorithm sorting your resume into a pile doesn’t know you. A friend who’s watched you work for years does. That gig is genuinely promising. It’s also a few hours a week, which is not exactly a foundation you build a retirement on, and not knowing whether it grows into something steadier or quietly evaporates is its own brand of restless. I’m hopeful and wary in roughly equal measure, which I’m told is just what it feels like right now for anyone looking for work.
Where that leaves me
I don’t have a clean ending for this one. I think the honest answer is that I’m doing all the things all at once. I’m still refining that workshop format and thinking about who’d actually show up to a class like this, because I believe there’s a real and maybe growing audience for “how to work with AI critically” rather than “how to code from scratch.” And I’m also thinking about picking up a paintbrush again, not as a backup plan, but as a way of staying sane while I figure out what a second act actually looks like at this stage.
If you’re in the same spot, somewhere past fifty, recently let go, watching the ladder you used to climb get dismantled rung by rung, I don’t have a tidy answer for you either. But I don’t think the answer is to pretend the old playbook still works. The playbook changed. The question is what we build with the part of it that’s still holds true.
As for the couch on the stairwell, I haven’t gotten it through yet. But I’ve stopped yelling “pivot” at it and started actually looking at the angles. I think progress looks more like that, than it does a breakthrough.
